filter_irrelevant_rules: Filter irrelevant decision rules using leave-one-out pruning

In CRE: Interpretable Discovery and Inference of Heterogeneous Treatment Effects

Filter irrelevant decision rules using leave-one-out pruning

Description

Filters the irrelevant decision rules. The irrelevant rules are interpreted as an error increase after removing a variable-value pair from the decision rules (see “Interpreting tree ensembles with the inTrees package” by Houtao Deng, 2019).

Usage

filter_irrelevant_rules(rules, X, ite, t_decay)

Arguments

rules	A list of rules.
X	A features matrix.
ite	An estimated ITE.
t_decay	The decay threshold for rules pruning.

Value

A list of “relevant” rules.

CRE documentation built on Oct. 19, 2024, 5:07 p.m.